Prompts that ship work.

Create five Meta ads for [PRODUCT] targeting [AUDIENCE]. For each ad give a hook, primary text (max 90 words), headline (max 40 chars) and description. Vary the angle: pain, aspiration, social proof, contrarian, curiosity. Output as a table.

Audit this landing page for a [AUDIENCE] buying [PRODUCT]. Score clarity, offer, proof, friction and CTA (1-5). Return the 3 highest-leverage changes with the rationale and expected impact on conversion rate.

Generate a positioning strategy for [PRODUCT]. Use April Dunford's structure: competitive alternatives, unique attributes, value themes, target segment, market category. Output as a one-page brief.

Write 5 cold email openers to [PERSONA] at [COMPANY_TYPE]. Each must be under 25 words, reference something specific and observable, and lead into a one-sentence value proposition. No flattery, no filler.

Break down the acquisition funnel for [PRODUCT]. For each stage (awareness → consideration → decision → activation) give the key metric, common leaks and the fastest experiment to run in week 1.

Given these customer insights: [PASTE]. Produce a creative brief with: single-minded proposition, 3 emotional angles, 3 rational angles, tone of voice guidance, and 6 concept directions with a 1-line description each.

Generate 20 scroll-stopping hooks for [PRODUCT] targeting [AUDIENCE]. Split evenly across: pattern interrupt, contrarian statement, hyper-specific claim, question, before/after. Each hook under 12 words.

Build an angle-testing matrix for [PRODUCT]. Rows: 5 pain points. Columns: 5 desired outcomes. Each cell = one testable ad concept with a hook, promise and proof element.

Write a UGC creator brief for [PRODUCT]. Include: hook (2 options), demo beats, key claims we can and cannot make, CTA line, technical requirements (aspect ratio, length, captions). Under 400 words.

Turn this hook into a 15-second video ad storyboard: [HOOK]. 6 frames. For each: on-screen text (max 8 words), b-roll direction, voiceover, sound cue.

Cluster these keywords into ad groups for [PRODUCT]: [KEYWORDS]. Optimize for intent tightness. For each cluster give: theme, 3 headlines, 2 descriptions, negative keyword suggestions.

Create an SEO content brief for the query '[QUERY]'. Include: search intent, top-ranking pattern, target word count, H2 outline, entities to cover, internal linking suggestions, and one differentiator we can own.

Write landing page copy for [PRODUCT] with this structure: hero (headline + subhead + CTA), 3 problem points, 3 solution points, 4 features, 3 objections handled, social proof block, FAQ (5 Qs), final CTA. Voice: [VOICE].

Rewrite the above-the-fold section for [PRODUCT]. Give 5 versions varying by clarity vs curiosity trade-off. Each version: headline, subheadline, CTA button. Score each 1-5 on clarity, specificity, differentiation.

For [PRODUCT] sold to [AUDIENCE], list the top 12 objections buyers have (price, trust, timing, fit, alternatives). For each: the raw quote we'd hear, the underlying belief, and one reframe that neutralizes it.

Write a 5-email welcome sequence for [PRODUCT]. Emails 1-5 goals: welcome + expectations, origin story, top use case, social proof, activation CTA. Each email <150 words. Include subject line + preheader.

Write a win-back email for churned customers of [PRODUCT]. Reference the specific reason for churn: [REASON]. Show what's changed. Include a low-friction offer. Max 120 words. No guilt, no discounts as the lead.

Audit the lifecycle emails for [PRODUCT]: [LIST]. Identify gaps in the customer journey (onboarding, activation, retention, expansion, win-back). For each gap: the missing trigger, the audience, the goal, and a 1-line brief.

Design an offer stack for [PRODUCT] at [PRICE]. Include: core outcome, 3 bonus items with a stated dollar value, urgency mechanic, risk reversal. Explain why each element neutralizes a specific objection.

We currently charge [PRICE] for [PRODUCT]. Propose 3 different pricing anchor strategies (decoy tier, alternative frame, unbundling). For each: the new price table, the customer psychology, and the risk.

For [PRODUCT] with [BUYER_PROFILE] and [SALES_MOTION], recommend free trial, freemium, both, or neither. Explain the trade-offs on CAC, activation, sales cycle and support load. Give the metrics we'd need to move the decision.

Design an event tracking schema for [PRODUCT]. List the 15-20 events we need across acquisition, activation and retention. For each: event name (snake_case), when it fires, key properties, and which report it feeds.

In plain English, write step-by-step instructions to build a GA4 exploration report that shows [METRIC] segmented by [DIMENSION] over the last 90 days, comparing paid vs organic. Include filters and calculated fields.

Compare last-click, data-driven and MMM for a [BUSINESS_TYPE] spending [BUDGET]/mo across [CHANNELS]. Explain what each will over- and under-credit, when to use which, and the minimum data volume required.

Propose a North Star metric for [PRODUCT]. Evaluate 3 candidates against: reflects value delivered, leads revenue, measurable weekly, movable by the team. Recommend one and 4 input metrics that drive it.

Design a 45-minute weekly growth review agenda for a [TEAM_SIZE] team on [PRODUCT]. Include: metrics reviewed, decision points, experiment updates, blockers. Assign owners and time per section.

Given these experiment ideas: [LIST]. Score each on impact, confidence, ease (1-10). Sort. For the top 3, add: hypothesis, primary metric, guardrail metric, minimum sample size, expected duration.

Here is a retention curve: [DATA]. Diagnose whether the product has a leaky bucket, an onboarding problem, or no product-market fit. Recommend the single next investigation to run and why.

Write a 30-minute customer interview script for users of [PRODUCT] who [BEHAVIOR]. Follow the Mom Test principles. 12 open questions structured around: context, current job, alternatives, tension, hopes.

Synthesize these 6 customer interviews: [TRANSCRIPTS]. Output: top 5 recurring themes with verbatim quotes, 3 tensions in the data, 3 signals of unmet need, one bet worth testing next.

Design a referral program for [PRODUCT] with [CURRENT_LTV]. Choose: reward type, reward timing, gating rules, share mechanics. Model expected referral rate and payback. Flag risks (fraud, cannibalization).

Our current copy sounds like [DESCRIPTION]. Rewrite this paragraph 4 ways: as Basecamp, as Linear, as Stripe, as Apple. Preserve meaning. After each version, note the trade-off in perceived audience.

Our [ACCOUNT] shows frequency [X] and rising CPM. Build a 30-day creative refresh plan: how many concepts per week, angle rotation, testing structure, and pause/scale rules based on [METRICS].

Propose 5 content pillars for a TikTok account for [PRODUCT] targeting [AUDIENCE]. For each pillar: theme, 5 hook formulas, 3 example video concepts, expected role in the funnel (top / mid / bottom).

Write a 60-second YouTube ad script for [PRODUCT]. Structure: 3s hook, problem, agitate, solution, proof, offer, CTA. Include voiceover, on-screen text and rough b-roll direction per beat.

Write a podcast guest pitch for [FOUNDER] to [SHOW]. Reference a recent episode. Propose 3 topic angles the host would want, tied to their audience. Under 120 words. No 'I love your show'.

Define the ICP for [PRODUCT]. Include: firmographics, technographics, trigger events, buying committee, budget range, disqualifiers. Then list 3 tightest ICP segments and score their reachability 1-5.

Write a 7-touch outbound sequence for [PERSONA] at [COMPANY_TYPE]. Mix email, LinkedIn, and one creative touch. Each email under 90 words. Include the trigger event and value hypothesis for the sequence.

Write a long-form sales page for [PRODUCT] at [PRICE]. Follow Eugene Schwartz's awareness stages. Include: lead, problem section, mechanism, offer, stack, guarantee, urgency, FAQ, PS.

Interview script to extract a case study from [CUSTOMER]. 15 questions structured around: their before state, why they chose us, first-week experience, specific results with numbers, quotable moments.

Turn this interview into a case study: [TRANSCRIPT]. Structure: challenge (100w), approach (200w), results with 3 numbers, quote block. Voice: understated, specific, no marketing adjectives.

Given this pixel setup: [CODE]. Identify: duplicate events, misfiring triggers, missing standard events, deduplication issues. Provide a fixed configuration and the QA steps to verify each event.

Design an MCP tool spec that lets Claude query [DATA_SOURCE]. Include: tool name, description, input schema, output schema, auth model, rate limit, and a natural language example call the model would make.

We're deploying an AI agent that can [ACTIONS] on [ACCOUNT]. Design guardrails: hard limits, require-approval thresholds, kill switches, drift detection, audit log format. Assume adversarial conditions.

Write a system prompt for an AI marketing assistant for [BRAND]. Encode: voice, no-go topics, factual guardrails, response formatting, escalation rules. Structure: role, capabilities, constraints, tone, examples.

This prompt is producing [BAD_OUTPUT]: [PROMPT]. Diagnose which prompt-engineering principles it violates. Suggest 3 rewrites that fix the issue, ranked by expected reliability. Explain the changes.

Turn this raw performance data into a stakeholder-ready weekly report: [DATA]. Sections: TL;DR (3 bullets), what worked, what didn't, next week's bets, risks. Tone: honest, no jargon, no filler.

For a RAG system indexing [CORPUS], rewrite this raw user question into 3 retrieval queries: [QUESTION]. Cover: keyword, semantic paraphrase, and decomposed sub-questions. Explain each rewrite briefly.

Generate 25 names for [PRODUCT]. Structure: 5 descriptive, 5 metaphorical, 5 invented, 5 founder-style, 5 category-defining. For the top 5, check .com availability heuristically and flag risks.

Write a brand voice guide for [BRAND] targeting [AUDIENCE]. Include: 3 voice attributes with 'we are / we are not' pairs, sentence-level do/don't examples, vocabulary list, edge cases (support, sales, error states).

Rewrite this changelog in a voice that feels like Linear: [CHANGELOG]. Keep it specific, human, and warm. Include emojis only if they replace words. Under 250 words per release.

Design a test plan for the [PRODUCT] homepage hero. Propose 4 headlines varying by: outcome vs mechanism vs identity vs contrarian. Define: primary metric, guardrails, sample size, expected duration.

Rewrite the pricing page for [PRODUCT]. Goals: clarify who each tier is for, kill decision fatigue, surface annual savings, handle 3 top objections. Include tier headers, feature list, FAQ (6 Qs), guarantee.

Design a post-purchase email flow for [PRODUCT]. Trigger, delay, subject, goal for each email. Include: order confirmation, unboxing, first-use nudge, review request, replenishment / expansion, VIP.

Rewrite these support macros for [BRAND]: [MACROS]. Keep the info, upgrade the voice to warm-professional. Add empathy openers, remove corporate hedging, make the next step unmissable.

Plan the launch of a community for [PRODUCT]. Platform decision, first 100 members strategy, week-1 content calendar, moderation rules, engagement rituals, success metrics at day 30/60/90.

Write the internal narrative for launching [FEATURE]. Include: the customer problem in their words, the shift this represents, why now, what we're deliberately not doing, and how we'll know it worked.

Design an ideal week for a founder of a [STAGE] company focused on [PRIORITY]. Block: deep work, sales calls, customer interviews, team, thinking. Include morning ritual, weekly review, and cut list.

Turn this customer review into 5 Meta ad copy variations: [REVIEW]. Preserve their exact language. Each variation: hook, primary text, headline. Keep the review's voice — don't clean it up.

Given this account snapshot: [DATA]. Diagnose the top 3 problems (structure, creative, targeting, tracking, bidding). For each: symptom, likely cause, 1-week fix, expected impact on CAC.

Design a taxonomy for a prompt library serving [TEAM]. Propose top-level categories (5-8), tags, metadata fields (model, input variables, output format, tested-on-date), and a naming convention for prompt IDs.

For this task: [TASK]. Compare a direct prompt vs a chain-of-thought prompt vs a decomposed multi-step chain. Predict output quality, cost, latency for each. Recommend one and justify.